SEARCH

SEARCH BY CITATION

Keywords:

  • aging;
  • cardiovascular disease;
  • inflammation;
  • lifestyle;
  • oxidative stress;
  • telomere

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

Evidence assembled over the last decade shows that average telomere length (TL) acts as a biomarker for biological aging and cardiovascular disease (CVD) in particular. Although essential for a more profound understanding of the underlying mechanisms, little reference information is available on TL. We therefore sought to provide baseline TL information and assess the association of prevalent CVD risk factors with TL in subjects free of overt CVD within a small age range. We measured mean telomere restriction fragment length of peripheral blood leukocytes in a large, representative Asklepios study cohort of 2509 community-dwelling, Caucasian female and male volunteers aged approximately 35–55 years and free of overt CVD. We found a manifest age-dependent telomere attrition, at a significantly faster rate in men as compared to women. No significant associations were established with classical CVD risk factors such as cholesterol status and blood pressure, yet shorter TL was associated with increased levels of several inflammation and oxidative stress markers. Importantly, shorter telomere length was associated with an increasingly unhealthy lifestyle, particularly in men. All findings were age and gender adjusted where appropriate. With these cross-sectional results we show that TL of peripheral blood leukocytes primarily reflects the burden of increased oxidative stress and inflammation, whether or not determined by an increasingly unhealthy lifestyle, while the association with classical CVD risk factors is limited. This further clarifies the added value of TL as a biomarker for biological aging and might improve our understanding of how TL is associated with CVD.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

The length of telomeres, the specialized terminal regions of chromosomes, from circulating peripheral blood leukocytes (PBL) has been proposed as a systemic index of biological aging. Human telomeres consist of tandem repeats of a noncoding TTAGGG hexamer (Blackburn, 2000). Telomere length (TL) is to a large extent genetically determined, yet in proliferating cells telomeric sequences are lost with each cell division due to the inability of DNA polymerase to complete the chromosome termini, which is reflected in an age-dependent telomere attrition (Lindsey et al., 1991; Blackburn, 2000; Graakjaer et al., 2006). In addition, in vitro experiments showed TL to be further modulated by environmental factors such as oxidative stress that induces direct DNA damage to the telomeric DNA (von Zglinicki et al., 1995). Furthermore, epidemiological results suggest that oxidative stress may be involved in in vivo telomere attrition (Demissie et al., 2006).

TL might therefore offer a valuable tool to measure biological aging and could in particular represent a helpful biomarker in cardiovascular disease (CVD) risk stratification, as CVD is considered as a typical aging disease (Bekaert et al., 2005a). In recent studies, telomere length was suggested to be associated with CVD and risk. Shorter telomeres were, for example, demonstrated in subjects who smoked (Nawrot et al., 2004; Valdes et al., 2005), had elevated pulse pressure (Jeanclos et al., 2000; Benetos et al., 2001), diabetes (Jeanclos et al., 1998), premature myocardial infarction (Brouilette et al., 2003), degenerative aortic valve stenosis (Kurz et al., 2006), chronic psychological stress (Epel et al., 2004, 2006), increased insulin resistance (Gardner et al., 2005), and an increased risk for CVD-induced mortality (Cawthon et al., 2003). While several of these results could not be confirmed by other studies, for example, the impact of smoking (Harris et al., 2006; Collerton et al., 2007; Fitzpatrick et al., 2007), the biomarker value of TL in cardiovascular aging itself was recently confirmed. In a male population aged between 45 and 65 years, statin-free subjects with shorter telomeres were at an apparent increased risk for coronary heart disease. In statin-treated subjects, however, this association was substantially attenuated (Brouilette et al., 2007). In an older population, considering both genders (approximately 65–90 years old), shorter TL length was found to be associated with an increased risk for stroke and myocardial infarction for subjects aged 73 or younger but not for their older counterparts (Fitzpatrick et al., 2007).

The question, however, remains whether shorter PBL TL at birth is a primary abnormality that reflects a predisposition to age-related diseases like CVD, thereby contributing to the genetic predisposition to CVD, or rather whether TL measured at any age is a mere reflection of the cumulative oxidative and inflammatory burden associated with the (cardiovascular) aging process, or whether both are involved.

In order to shed light on these issues and unravel the mechanisms by which TL might be associated with increased CVD risk, the present study assessed TL within a longitudinal population study, Asklepios. This study was designed to investigate the interplay between aging, cardiovascular hemodynamics and inflammation in (preclinical) CVD. The first round of the longitudinal Asklepios study is unique in that it is the first time TL has been studied in a large cohort of subjects free of overt CVD and not only for both sexes but also within a small age range, hence excluding possible generation-dependent confounding effects. The Asklepios volunteers are at lower immediate CVD risk compared to older subjects. Although the current risk assessment might therefore be somewhat less sensitive, our experimental set-up has the advantage of being less biased by characteristics associated with CVD, as opposed to real risk factors (Rietzschel et al., 2007).

Here we report the baseline values for TL in the Asklepios study population and evaluate the presumed associations between PBL TL and the most important cardiovascular risk factors, that is, whether TL is determined by (i) age and gender, (ii) other classical risk factors, (iii) nonclassical risk factors, and (iv) lifestyle, including readily modifiable risk factors. As all subjects were free of overt CVD, our conclusions are also valid in a general aging context.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

General considerations and baseline TL characteristics

The Asklepios cohort comprises 2524 subjects aged approximately 35–55 years (Rietzschel et al., 2007). In ten cases, reliable mean telomere restriction fragment (TRF) length measurements could not be obtained due to bad sample quality (blood or DNA) or technical problems. The resulting 2514 samples include 1293 women and 1221 men. From the total population, 2359 duplicate mean TRF assessments were available. Optimal conditions during sample handling and strictly following the improved protocol allowed an interassay coefficient of variation of 2.76 ± 2.20% (correlation: R = 0.867). No significant differences in mean TRF were observed for those samples for which only single measurements were available as compared to duplicate ones (P = 0.590). Five outliers were excluded based on visual examination. The resulting population under study thus consisted of 2509 subjects (1291 women and 1218 men), including 74 participants older than 55 and 15 younger than 35. The average mean TRF within this population was measured at 7.873 kb (standard deviation, SD, 0.728 kb). The baseline characteristics of this study population are summarized in Table 1. A detailed overview (for the 2524 subjects) has already been reported (Rietzschel et al., 2007).

Table 1.  Baseline characteristics summary of the 2509 subjects from the Asklepios study included in the analysis
 Men (n = 1218) DescriptivesWomen (n = 1291) DescriptivesP value*Total (n = 2509) Descriptives
  • Descriptives are mean ± standard deviation (SD) or median [interquartile range] where appropriate.

  • NSAID, nonsteroidal anti-inflammatory drugs; ASA, acetylsalicylic acid.

  • *

    Independent samples t-test (except for

  • – Mann–Whitney U-test) to compare men and women.

  • Pearson χ2-test to compare men and women.

Age (year)46.1 ± 5.945.9 ± 6.00.31746.0 ± 6.0
Systolic blood pressure (mmHg)130.7 ± 12.7123.3 ± 14.5< 0.001126.9 ± 14.1
Diastolic blood pressure (mmHg)82.4 ± 9.877.8 ± 9.7< 0.00180.1 ± 10.0
Triglycerides (mg dL−1)104 [75.3–156]79.2 [59.7–112]< 0.00190.7 [65.4–131.0]
Total cholesterol (mg dL−1)219.1 ± 37.7214.2 ± 35.5< 0.001216.6 ± 36.7
LDL cholesterol (mg dL−1)137.4 ± 34.3124.9 ± 32.9< 0.001130.9 ± 34.2
HDL cholesterol (mg dL−1)56.0 ± 13.870.6 ± 17.1< 0.00163.5 ± 17.2
oxLDL (U L−1)100.7 ± 38.891.6 ± 38.4< 0.00196.0 ± 38.9
hs-CRP (mg L−1)1.05 [0.56–2.04]1.42 [0.62–3.42]< 0.0011.2 [0.59–2.70]
IL-6 (pg mL−1)0.79 [0.00–1.60]0.75 [0.00–1.50]0.1930.78 [0.00–1.54]
Serum uric acid (mg dL−1)6.06 ± 1.274.32 ± 1.05< 0.0015.16 ± 1.45
Fibrinogen (mg dL−1)313.7 ± 59.4335.9 ± 64.9< 0.001325.1 ± 63.2
BMI (kg m−2)26.5 ± 3.725.1 ± 4.6< 0.00125.8 ± 4.3
Waist circumference (cm)93.8 ± 10.480.5 ± 11.3< 0.00186.9 ± 12.7
 Number (%)Number (%)P valueNumber (%)
Smoking status  < 0.001 
 Never504 (41.4)781 (60.5) 1285 (51.2)
 Former419 (34.4)280 (21.7) 699 (27.9)
 Current295 (24.2)230 (17.8) 525 (20.9)
Presence of hypertension412 (33.8)312 (24.2)< 0.001724 (28.9)
Antihypertensive therapy118 (9.7)145 (11.2)0.207263 (10.5)
NSAID/ASA therapy47 (3.9)69 (5.3)0.076116 (4.6)

TL and classical CVD risk factors age and gender (Fig. 1, Table 2)

image

Figure 1. Telomere length and gender. Mean telomere restriction fragment (TRF) length is shorter in men (—◊—) compared to women (—▴—)

Download figure to PowerPoint

Table 2.  Reference values for telomere length (TL) for different ages and genders
Age, yearsMenWomen
nMean (bp)SD (bp)nMean (bp)SD (bp)
  • SD, standard deviation.

  • *

    Including 15 participants younger than 35 years but older than or equal to 30 years.

  • Including 74 participants older than 55 years but younger than 60 years.

Younger than 41*29580150.7793358088701
41–4531678510.6553218049743
46–5030577040.6823087909710
Older than 5130275720.6603277775734
All121877850.71312917956732

Mean TRF was shown to be inversely correlated with age, accounting for 4.3% of the variance (P < 1E-25). The yearly telomere attrition calculated from these cross-sectional baseline data was estimated at 26 base pairs (bp).

Mean TRF was found to be dependent on gender (P < E-8) with men having on average 172 bp shorter telomeres than women in this middle-aged cohort. After age adjustment, this difference was 166 bp (P < E-8). More importantly, our cross-sectional data show that telomere attrition proceeds faster in men (30.0 bp per year, R2 = 0.062) compared to women (20.3 bp per year, R2 = 0.028). This was statistically confirmed with a general linear model (GLM) containing both age, gender and the age–gender interaction term: the latter remained significant (P = 0.041) while gender did not (P = 0.205). Since this underscores a gender-dependent TL attrition rate, the dataset was split according to gender for further analysis. Only when the association between TL and a certain characteristic was (at least borderline) significant in both genders, was a GLM combining both genders created. In these GLMs, gender adjustments were performed using the age–gender interaction term.

TL and other not readily modifiable classical CVD risk factors (Table 3)

Table 3.  Association between telomere length (TL) and classical cardiovascular disease (CVD) risk factors
 MenWomen
β coeff. (bp)R2 (P)β coeff. (bp)R2 (P)
  • β coeff., unstandardized β coefficient; Adj. diff., age-adjusted difference; ln, natural logarithm.

  • *

    After exclusion of subjects taking antihypertensive therapy.

  • Including subjects under current antihypertensive therapy.

Total cholesterol (mg dL−1)–0.171< 0.001 (0.748)0.115< 0.001 (0.844)
LDL-C (mg dL−1)–0.071< 0.001 (0.931)0.258< 0.001 (0.695)
HDL-C (mg dL−1)–2.0700.002 (0.149)0.547< 0.001 (0.641)
ln triglycerides (mg dL−1)0.1910.001 (0.208)0.107< 0.001 (0.827)
Glycemia (mg dL−1)0.398< 0.001 (0.846)–2.5150.001 (0.293)
Systolic blood pressure (mmHg)*1.5170.001 (0.344)–0.359< 0.001 (0.803)
Diastolic blood pressure (mmHg)*1.7360.001 (0.404)–0.482< 0.001 (0.819)
 Mean (bp) (n)Adj. diff. (P)Mean (bp) (n)Adj. diff. (P)
Blood pressure status
 Normotensive7801 (806)07987 (979)0
 Hypertensive7753 (412)–20 (0.631)7860 (312)–64 (0.186)

No significant associations with TL were found for total, low-density lipoprotein (LDL), nor high-density lipoprotein (HDL) cholesterol. Neither were there significant associations between TL and log-transformed serum triglycerides nor glucose concentrations, respectively. Hypertensive subjects did not exhibit shorter telomeres than their normotensive counterparts. After exclusion of antihypertensive treated subjects, there do not appear to be any significant associations between mean TRF and blood pressure (BP, systolic and diastolic). The readily modifiable risk factors such as body mass index (BMI) and smoking status will be discussed in the context of lifestyle-related risk factors (cf. infra).

TL and novel CVD risk factors (Table 4)

Table 4.  Association between telomere length (TL) and inflammation, oxidative stress markers, respectively, and other recently proposed risk factors
 MenWomen
β coeff. (bp)R2 (P)β coeff. (bp)R2 (P)
  • β coeff., unstandardized β coefficient; Adj. diff., age-adjusted difference; ln, natural logarithm.

  • *

    Significant at the 0.05 level.

  • Significant at the 0.01 level.

OxLDL (U L−1)–1.0040.003 (0.051)–1.1780.004 (0.026)*
ln hs-CRP (mg L−1)–41.7780.004 (0.029)*–34.4760.003 (0.045)*
Fibrinogen (mg dL−1)–0.9170.006 (0.007)–0.166< 0.001 (0.600)
Uric acid (mg dL−1)–50.4580.009 (0.001)–35.5160.003 (0.068)
 Mean (bp) (n)Adj. diff. (P)Mean, bp (n)Adj. diff. (P)
IL-6
 Lowest half7851 (611)07998 (646)0
 Highest half7718 (607)–115 (0.004)7915 (645)–72 (0.075)
Uric acid status
 Normal7835 (800)07976 (1210)0
 Hyperuricemia7689 (418)–119 (0.004)7658 (81)–260 (0.002)

The associations between mean TRF and several more recently suggested CVD risk factors were evaluated, except for those largely related with lifestyle (cf. infra).

No age-adjusted associations between mean TRF values and heart rate, total plasma homocysteine (tHCY), and creatinine level were observed (data not shown, all P > 0.25).

The inflammation markers high-sensitive C-reactive protein [hs-CRP, ln transformed; men: P = 0.029; women: P = 0.045 to P = 0.018 when adjusted for current use of hormonal contraceptives as these increase hs-CRP levels (Frohlich et al., 1999)] and fibrinogen (only in men: P = 0.007) showed a negative association with age-adjusted TL. The specific nature of interleukin 6 (IL-6) measurements (highly skewed, zero rich) does not allow a linear statistical approach and the results were therefore dichotomized at the median (cut-off in men: 0.800; women: 0.760). Age-corrected TL was 72 bp shorter in women with high IL-6 concentrations compared to low IL-6 concentrations (P = 0.075). In men, a significant 115-bp difference was noted (P = 0.004). Oxidized LDL (oxLDL) showed a negative association with age-adjusted TL (borderline in men: P = 0.051; women: P = 0.026). In men, an apparent negative age-adjusted association between mean TRF and serum uric acid was found. Comparing normal men and women with their hyperuricemic counterparts showed 260-bp (P = 0.002) and 119-bp (P = 0.004) longer telomeres in normal women and men, respectively.

Both hs-CRP, elevated IL-6, oxLDL and serum uric acid concentration appear to play a role in both genders. In GLMs adjusting for age and gender all covariates/factors were clearly significant (for all: P < 0.005) when merging both genders in one dataset, suggesting an impact of inflammation and oxidative stress on the general population. Adjusting for gender as such or for gender–age interaction did not significantly alter our findings, nor did adjustment for the use of particular medications [statins, acetylsalicylic acid-based medication (ASA)/nonsteroidal anti-inflammatory drugs (NSAID)].

TL and lifestyle-related risk factors (Table 5)

Table 5.  Association between telomere length (TL) and anthropometric factors and CVD risk elevating lifestyle, respectively
 MenWomen
β coeff. (bp)R2 (P)β coeff.R2 (P)
  • β coeff., unstandardized β coefficient; Adj. diff., age-adjusted difference; ln, natural logarithm.

  • *

    Significant at the 0.05 level.

Height (cm)–2.099< 0.001 (0.498)–2.662< 0.001 (0.441)
Weight (kg)–1.6300.001 (0.306)–1.9880.001 (0.213)
BMI (kg m−2)–4.4940.001 (0.406)–4.1220.001 (0.351)
Waist circumference (cm)–3.4330.003 (0.077)–2.3340.001 (0.202)
ln physical activity (MET × times per week)25.2100.002 (0.152)–3.306< 0.001 (0.806)
ln alcohol consumption (units alcohol per week)–8.033< 0.001 (0.630)–7.453< 0.001 (0.682)
ln smoking quantity (pack-year)–20.4210.002 (0.164)5.742< 0.001 (0.753)
Fruit and vegetable (g day−1)  0.042< 0.001 (0.751) 0.080< 0.001 (0.526)
Unhealthy lifestyle–19.4480.004 (0.033)*–12.9170.001 (0.197)
 Mean (bp) (n)Adj. diff. (P)Mean (bp) (n)Adj. diff. (P)
BMI status All: P = 0.619 All: P = 0.426
 Normal7851 (427)07995 (759)0
 Overweight7751 (580)–43 (0.332)7889 (357)–61 (0.192)
 Obese7744 (211)–33 (0.579)7928 (175)–21 (0.734)
Smoking status All: P = 0.231 All: P = 0.527
 Never7833 (504)07950 (781)0
 Former7764 (419)12 (0.800)7987 (280)53 (0.292)
 Active7732 (295)–73 (0.151)7943 (230)–8 (0.886)
TL and individual lifestyle-related risk factors

We assessed the impact of CVD-related lifestyle characteristics on mean TRF. Due to their close connection with lifestyle parameters, we decided to additionally include anthropometric parameters.

After age adjustment, the classical CVD risk factor BMI no longer demonstrated a significant association with mean TRF. As for BMI, no significant age-adjusted associations with waist-to-hip ratio (WHR) or directly measured anthropometric properties were obtained. TL was not significantly associated with log-transformed physical activity. Furthermore, TL was not significantly different between active, former and never smokers, nor between ever and never smokers. After exclusion of former smokers, no significant difference between never smokers and active smokers was observed. Log-transformed number of pack-years was not significantly associated with TL after age-adjustment.

In neither male nor female subjects was the age-adjusted association between TL and log transformed weekly alcohol consumption found to be significant. TL was also not significantly associated with the daily intake of fruit and vegetables (g day−1).

TL and increasingly unhealthy lifestyle

Although never significant, the directions of the associations between TL and the individual lifestyle variables often suggested an unhealthy lifestyle to have a negative impact on TL. Furthermore, in an exploratory analysis, subjects with a fast weight gain since the age of 18 had shorter telomeres (P = 0.029, age and gender adjusted), although this was not significant for the separate genders (P = 0.122 and P = 0.104 for men and women, respectively). This paradigm also suggests a lifestyle effect. As the associations between TL and the individual lifestyle components might be obscured by the complexity of their interactions, for example, current smokers are generally leaner than their currently nonsmoking counterparts (P = 0.001, age and sex adjusted), we tried to assess the impact of a generally unhealthy lifestyle on TL. Therefore, a novel composite unhealthy lifestyle variable was constructed based on the individual lifestyle components.

The variables waist circumference and pack-years smoking were selected, together with weekly alcohol intake, fruit and vegetable intake and physical activity. The lifestyle variables were divided into tertiles (separately for each gender), which were used to calculate the unhealthy lifestyle score (cf. Experimental procedures).

Despite the lack of association of the individual variables with TL and the very conservative way of selecting the variables and cut-offs, TL was significantly associated with an increasingly unhealthy lifestyle (P = 0.013, age and gender adjusted). When looking at the genders separately, the association was significant in men (P = 0.033), but not in women (P = 0.197).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

Although recent studies strongly support the biomarker value of TL in biological aging and CVD in particular, the underlying mechanisms are largely unknown. The relatively young Asklepios study population free of overt CVD provides an ideal background population to test how TL might be associated with aging in general and CVD risk in particular.

The hypothesis of TL acting as a biological aging biomarker is based on a number of crucial observations. First, TL is shown to shorten with chronological calendar age. We have confirmed this age-dependent telomere attrition. This attrition is characterized by sexual dimorphism, in accordance with other studies (Benetos et al., 2001; Okuda et al., 2002; Nawrot et al., 2004; Bischoff et al., 2005; Nordfjall et al., 2005; Unryn et al., 2005; Mayer et al., 2006). This indicates that in both genders, different mechanisms might have different implications for telomere erosion. To the best of our knowledge, this is the first time that evidence of a gender-dependent telomere attrition rate could be statistically validated in a young to middle-aged population. Furthermore, these associations provide a strong positive control for the sensitivity of our methodology.

These associations also support the possible biomarker point of view as they parallel the age and gender risk factors for CVD. To our surprise, however, other classical cardiovascular risk factors, such as BMI, smoking status, cholesterol fraction concentrations and blood pressure, were at best only moderately associated with shorter PBL telomeres. On the other hand, we were able to establish a link between TL and oxidative stress/inflammation.

At the systemic level, the biomarker value of TL partially lies in its capability to reflect the cumulative impact of increased cell turnover. CVD is considered to be a chronic inflammatory disease, and therefore elevated serum levels of the markers fibrinogen, hs-CRP and IL-6 are presumed indicators of a higher CVD risk (Lobbes et al., 2006). The consistent pattern of negative associations between inflammatory markers and TL as observed in our study further strengthens the hypothesis that the increased cell turnover due to inflammation is reflected in an increased telomere attrition and therefore shorter baseline TL, thus providing a direct link between CVD and TL. This link had already been suggested by Aviv et al. who described a negative correlation with CRP in women aged under 50 (2006) and by Fitzpatrick et al. who found TL to be associated with IL-6 and CRP, albeit only for men or subjects aged 73 years or younger (2007). These results suggest an age dependency on the association between TL and CRP, which might explain the lack of significance when tested in an 85-year-old population (Collerton et al., 2007).

Supplementary strength of TL as a biological aging biomarker is derived from the fact that the environmental context, that is, the chain of oxidative damage but also the physiological response, influences the rate of telomere attrition at the systemic level. Oxidative stress has indeed been associated with accelerated telomere attrition in vivo (Demissie et al., 2006). The observed association between baseline TL and serum oxLDL in our study is therefore consistent with an effect of oxidative stress on telomere shortening. Furthermore, we found telomeres to be shorter in subjects with higher serum uric acid concentrations. Although the specific role of uric acid is controversial, current evidence shows that increased concentrations are associated with oxidative stress-related increased CVD risk (Johnson et al., 2003). Based on in vitro experiments, Xu and co-workers showed that telomere attrition was increased by homocysteine (2000). Hyperhomocysteinemia is a presumed marker for increased CVD risk, probably through a mechanism of increased oxidative stress (Weiss, 2005). As the specific role of homocysteine (but also, e.g. serum uric acid) in CVD is still to be elucidated, it is very difficult to pinpoint a possible reason for the lack of association between TL and homocysteine in the Asklepios population.

The major advantage of this study is the fact that the study population is a large, young to middle-aged population, free of overt CVD, sampled within a small age range. This avoids the occurrence of generation dependent, and other, confounding effects, but is also a limitation for the study as it reduces the capacity to discriminate long-term effects, for example, the impact of smoking status. This might explain the lack of certain significant associations compared to other reports.

Some groups, for example, were able to link shorter telomeres with hypertension (Jeanclos et al., 2000; Benetos et al., 2001, 2004; Demissie et al., 2006), an important cardiovascular risk factor (Palatini & Julius, 1997). In a cohort of women aged 18–79 years, associations of TL with BMI-defined obesity and smoking status have also been shown (Valdes et al., 2005). We suggest that these associations were not found in our study due to the study design. This is supported by the fact that in recent reports from other studies in a narrow-age range the risk factor value of TL for CVD could be established, but no significant associations with smoking status or hypertension could be found (Brouilette et al., 2007; Fitzpatrick et al., 2007). Our results therefore imply that the primary effects of CVD risk factors on TL are principally associated with inflammation and oxidative stress.

Interestingly, we found a relationship between telomere erosion and a proaging lifestyle, particularly in men, defined on the basis of increased smoking activity, waist circumference, and alcohol consumption, and decreasing physical activity and fruit and vegetable intake. This association was not significant in women, which might be attributed to their generally healthier lifestyle (data not shown). Although each single habit had only a subtle contribution, the total impact increased almost linearly for each added proaging lifestyle aspect. Notwithstanding the lack of statistical evidence for the individual components (the fact that all of these aspects have been suggested as CVD risk factors might be of interest). This unhealthy lifestyle is therefore most likely an important contributor to the oxidative stress/inflammation-induced telomere attrition. This is supported by the observation that several of the involved oxidative stress and inflammation parameters are significantly associated with the unhealthy lifestyle score (data not shown), causing less but still significant associations between each parameter and TL after adjustment for lifestyle. These conclusions all remained valid after exclusion of the subjects for which only one TL measurement was available.

Altogether, the results of this first round of the Asklepios study show that chronic oxidative stress and inflammation, at least partially, explain how CVD risk is linked with increased telomere attrition. The fact that TL is not associated with other, classical, CVD risk factors suggests an added value for TL as a marker for biological aging and CVD in particular. Furthermore, its capacity to register the cumulative impact of inflammation and oxidative stress during lifetime might surpass the ability of other biomarkers that are point estimates and only monitor these processes at a given moment in time. This hypothesis also implies a biomarker role for the telomere attrition rate itself next to baseline TL.

The predictive power of PBL TL or attrition rate as systemic biomarkers further awaits the results of the longitudinal follow-up rounds of the Asklepios study. This aspect is comprised within the outline and objectives in the follow-up rounds of the longitudinal study.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

The detailed protocols are contained in the Asklepios study methodological article, which focuses on the rationale, design, methods, exclusion criteria and baseline characteristics of the Asklepios study (Rietzschel et al., 2007). Relevant information has been summarized below.

Study subjects

Study subjects were drawn from a pool of 8121 subjects aged 35–55 years and living in the twinned Belgian communities of Erpe-Mere and Nieuwerkerken (Flanders, Belgium). Volunteering partners of the study subjects were also examined and included in the study if they were aged between 30 and 60 years. Altogether, we enrolled a representative cohort of 2524 subjects during a 2-year inclusion period that started in October 2002 and ended in September 2004.

This study complies with the Declaration of Helsinki. The ethical committee of the Ghent University Hospital approved the research protocol, and written informed consent was obtained from each participant prior to enrollment in the study.

Screening and participant examination

The primary inclusion point of the study was a group of 89 general physicians (see Supplementary Appendix), who reviewed a detailed personal medical history, conventional cardiovascular risk profile and an overview of drug use and referred candidates to a single cardiologist who performed all measurements single center, single device, and in a fixed order. All subjects had fasted and refrained from smoking for at least 6 h prior to examination. Participants with a recent active infection/inflammation were asked to return for blood sampling when their symptoms had decreased (> 10 days).

A study questionnaire on smoking history, weekly alcohol consumption, fruit and vegetable intake and physical activity was completed. All sources of tobacco were included in the smoking-related analyses. Log-transformed number of pack-years was calculated as ln(pack-years + 1) to include never smokers. The same zero-filling procedure was performed to calculate the log-transformed alcohol consumption. Physical activity was defined as the metabolic equivalent (MET) of the activity multiplied by the number of times per week this activity was carried out, and was also zero-filled and log-transformed to obtain the log-transformed physical activity.

Study components: basic clinical and anthropometric data

Body size parameters (height, weight, waist and hip circumference) were measured. BMI and WHR were calculated. Fast weight gain was defined as an average annual increase in weight of more than 0.5 kg since the age of 18. BP was recorded using bilateral brachial sphygmomanometric measurements in triplicate with 1-min intervals on a rested participant (blinded to the analysis results) and averaged. Subjects actively treated with antihypertensive drugs or exhibiting a systolic BP larger than or equal to 140 mmHg and/or a diastolic BP larger than or equal to 90 mmHg were designated as being hypertensive.

Study components: biochemical data

Fasting whole blood ethylenediamine tetraacetic acid (EDTA), citrate plasma and serum aliquots were collected, cooled at 4 °C and employed for routine analyses in a reference laboratory (Laboratory for Clinical Biology, UZ-Ghent; ISO 9002; NBN EN 45001) according to an ISO 17025 Beltest accreditation. Serum glucose, creatinine, total and HDL cholesterol, triglyceride uric acid and plasma fibrinogen concentrations were measured using commercial reagents (Roche Diagnostics, Mannheim, Germany). The cut-off for hyperuricemia was defined at 6 mg dL−1 for women and 6.5 mg dL−1 for men. LDL cholesterol concentrations were calculated using the Friedewald formula (Friedewald et al., 1972).

Aliquots for inflammation and oxidative stress assessments were cooled at –20 °C prior to analysis. hs-CRP concentrations were measured by high-sensitive immunoturbidimetry (Roche Diagnostics). Serum IL-6 was measured by a chemiluminescent immunometric assay (DPC, Los Angeles, CA, USA). oxLDL was assayed in serum using an mAb-4E6-based sandwich enzyme-linked immunosorbent assay (ELISA) technique (Mercodia, Uppsala, Sweden). tHCY was measured by a fluorescence polarization immunoassay (Abbott, IL, USA).

Study components: mean leukocytes TRF analysis

DNA isolation was performed maximally 3 days after blood sampling. High-quality genomic DNA samples were extracted from whole blood EDTA using the Puregene™ Genomic purification kit (Gentra Systems, Minneapolis, MN, USA). The yield and quality of the duplicate DNA samples were assessed and samples were stored at –80 °C. Aliquots of 5 µg were short-term stored at –20 °C for TRF analysis.

We performed a standard TRF methodology for telomere sizing but with inclusion of some in-house modifications (Bekaert et al., 2005b). TLs were measured as weighted mean TRF values. Duplicate measurements were averaged. Telomere attrition was here defined as the yearly loss in TL as estimated by the cross-sectional data. TLs were experimentally determined while blinded for subjects ID.

Construction of the unhealthy lifestyle score

For each subject the unhealthy lifestyle score was calculated as the sum of the tertiles of five individual lifestyle variables (waist circumference, daily fruit and vegetable intake, physical activity, weekly alcohol intake and pack-years smoking), which resulted in a score between 5 and 15. The lowest tertile of smoking quantity consisted of all never smokers. The score was calculated so that subjects belonging to the lowest tertiles of physical activity and fruit and vegetables intake, and to the highest tertiles of pack-years smoking, weekly alcohol intake and waist circumference were assigned the largest unhealthy lifestyle score.

Statistical analysis

Statistical analysis was performed with SPSS 14.0.1 (SPSS Inc., Chicago, IL, USA). Level of significance was 0.05. Standard statistical methods were employed, including GLM, the independent t-test, the Mann–Whitney U-test and Pearson's χ2-test. Unless explicitly mentioned otherwise, each statistical model adjusted for age and contained mean TRF as the dependent variable. Gender-combining models adjusted for age and gender. For GLMs P values and partial eta-squared values (indicated as R2s) were reported.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References

The authors are grateful to Fien De Block, Dimitri Broucke, Sofie De Schynkel and Frida Brusselmans for technical assistance. We also would like to thank Professor Jan Tavernier from the Biochemistry Department of Ghent University Hospital for the use of laboratory facilities. This article is supported by grants from the National Fund for Scientific Research – Vlaanderen (G.0427.03). Sofie Bekaert and Tim De Meyer were funded by grants from the Special Research Fund of Ghent University (grant 01109502 and 011D10004, respectively).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. References
  • Aviv A, Valdes A, Gardner JP, Swaminathan R, Kimura M, Spector TD (2006) Menopause modifies the association of leukocyte telomere length with insulin resistance and inflammation. J. Clin. Endocrinol. Metab. 91, 635640.
  • Bekaert S, De Meyer T, Van Oostveldt P (2005a) Telomere attrition as ageing biomarker. Anticancer Res. 25, 30113021.
  • Bekaert S, Van Pottelbergh I, De Meyer T, Zmierczak H, Kaufman JM, Van Oostveldt P, Goemaere S (2005b) Telomere length versus hormonal and bone mineral status in healthy elderly men. Mech. Ageing Dev. 126, 11151122.
  • Benetos A, Gardner JP, Zureik M, Labat C, Xiaobin L, Adamopoulos C, Temmar M, Bean KE, Thomas F, Aviv A (2004) Short telomeres are associated with increased carotid atherosclerosis in hypertensive subjects. Hypertension 43, 182185.
  • Benetos A, Okuda K, Lajemi M, Kimura M, Thomas F, Skurnick J, Labat C, Bean K, Aviv A (2001) Telomere length as an indicator of biological aging: the gender effect and relation with pulse pressure and pulse wave velocity. Hypertension 37, 381385.
  • Bischoff C, Graakjaer J, Petersen HC, Jeune B, Bohr VA, Koelvraa S, Christensen K (2005) Telomere length among the elderly and oldest-old. Twin Res. Hum. Genet. 8, 425432.
  • Blackburn EH (2000) Telomere states and cell fates. Nature 408, 5356.
  • Brouilette SW, Moore JS, McMahon AD, Thompson JR, Ford I, Shepherd J, Packard CJ, Samani NJ (2007) Telomere length, risk of coronary heart disease, and statin treatment in the West of Scotland Primary Prevention Study: a nested case-control study. Lancet 369, 107114.
  • Brouilette S, Singh RK, Thompson JR, Goodall AH, Samani NJ (2003) White cell telomere length and risk of premature myocardial infarction. Arterioscler. Thromb. Vasc. Biol. 23, 842846.
  • Cawthon RM, Smith KR, O’Brien E, Sivatchenko A, Kerber RA (2003) Association between telomere length in blood and mortality in people aged 60 years or older. Lancet 361, 393395.
  • Collerton J, Martin-Ruiz C, Kenny A, Barrass K, Von Zglinicki T, Kirkwood T, Keavney B (2007) Telomere length is associated with left ventricular function in the oldest old: the Newcastle 85+ study. Eur. Heart J. 28, 172176.
  • Demissie S, Levy D, Benjamin EJ, Cupples LA, Gardner JP, Herbert A, Kimura M, Larson MG, Meigs JB, Keaney JF, Aviv A (2006) Insulin resistance, oxidative stress, hypertension, and leukocyte telomere length in men from the Framingham Heart Study. Aging Cell 5, 325330.
  • Epel ES, Blackburn EH, Lin J, Dhabhar FS, Adler NE, Morrow JD, Cawthon RM (2004) Accelerated telomere shortening in response to life stress. Proc. Natl Acad. Sci. USA 101, 1731217315.
  • Epel ES, Lin J, Wilhelm FH, Wolkowitz OM, Cawthon R, Adler NE, Dolbier C, Mendes WB, Blackburn EH (2006) Cell aging in relation to stress arousal and cardiovascular disease risk factors. Psychoneuroendocrinology 31, 277287.
  • Fitzpatrick AL, Kronmal RA, Gardner JP, Psaty BM, Jenny NS, Tracy RP, Walston J, Kimura M, Aviv A (2007) Leukocyte telomere length and cardiovascular disease in the cardiovascular health study. Am. J. Epidemiol. 165, 1421.
  • Friedewald WT, Levy RI, Fredrickson DS (1972) Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin. Chem. 18, 499502.
  • Frohlich M, Doring A, Imhof A, Hutchinson WL, Pepys MB, Koenig W (1999) Oral contraceptive use is associated with a systemic acute phase response. Fibrinolysis Proteolysis 13, 239244.
  • Gardner JP, Li S, Srinivasan SR, Chen W, Kimura M, Lu X, Berenson GS, Aviv A (2005) Rise in insulin resistance is associated with escalated telomere attrition. Circulation 111, 21712177.
  • Graakjaer J, Der-Sarkissian H, Schmitz A, Bayer J, Thomas G, Kolvraa S, Londono-Vallejo JA (2006) Allele-specific relative telomere lengths are inherited. Hum. Genet. 119, 344350.
  • Harris SE, Deary IJ, MacIntyre A, Lamb KJ, Radhakrishnan K, Starr JM, Whalley LJ, Shiels PG (2006) The association between telomere length, physical health, cognitive ageing, and mortality in non-demented older people. Neurosci. Lett. 406, 260264.
  • Jeanclos E, Krolewski A, Skurnick J, Kimura M, Aviv H, Warram JH, Aviv A (1998) Shortened telomere length in white blood cells of patients with IDDM. Diabetes 47, 482486.
  • Jeanclos E, Schork NJ, Kyvik KO, Kimura M, Skurnick JH, Aviv A (2000) Telomere length inversely correlates with pulse pressure and is highly familial. Hypertension 36, 195200.
  • Johnson RJ, Kang DH, Feig D, Kivlighn S, Kanellis J, Watanabe S, Tuttle KR, Rodriguez-Iturbe B, Herrera-Acosta J, Mazzali M (2003) Is there a pathogenetic role for uric acid in hypertension and cardiovascular and renal disease? Hypertension 41, 11831190.
  • Kurz DJ, Kloeckener-Gruissem B, Akhmedov A, Eberli FR, Buhler I, Berger W, Bertel O, Luscher TF (2006) Degenerative aortic valve stenosis, but not coronary disease, is associated with shorter telomere length in the elderly. Arterioscler. Thromb. Vasc. Biol. 26, e114e117.
  • Lindsey J, McGill NI, Lindsey LA, Green DK, Cooke HJ (1991) In vivo loss of telomeric repeats with age in humans. Mutat. Res. 256, 4548.
  • Lobbes MB, Lutgens E, Heeneman S, Cleutjens KB, Kooi ME, Van Engelshoven JM, Daemen MJ, Nelemans PJ (2006) Is there more than C-reactive protein and fibrinogen? The prognostic value of soluble CD40 ligand, interleukin-6 and oxidized low-density lipoprotein with respect to coronary and cerebral vascular disease. Atherosclerosis 187, 1825.
  • Mayer S, Bruderlein S, Perner S, Waibel I, Holdenried A, Ciloglu N, Hasel C, Mattfeldt T, Nielsen KV, Moller P (2006) Sex-specific telomere length profiles and age-dependent erosion dynamics of individual chromosome arms in humans. Cytogenet. Genome Res. 112, 194201.
  • Nawrot TS, Staessen JA, Gardner JP, Aviv A (2004) Telomere length and possible link to X chromosome. Lancet 363, 507510.
  • Nordfjall K, Larefalk A, Lindgren P, Holmberg D, Roos G (2005) Telomere length and heredity: indications of paternal inheritance. Proc. Natl Acad. Sci. USA 102, 1637416378.
  • Okuda K, Bardeguez A, Gardner JP, Rodriguez P, Ganesh V, Kimura M, Skurnick J, Awad G, Aviv A (2002) Telomere length in the newborn. Pediatr. Res. 52, 377381.
  • Palatini P, Julius S (1997) Heart rate and the cardiovascular risk. J. Hypertens. 15, 317.
  • Rietzschel ER, De Buyzere ML, Bekaert S, Segers P, De Bacquer D, Cooman L, Van Damme P, Cassiman P, Langlois M, Van Oostveldt P, Verdonck P, De Backer G, Gillebert TC (2007) Rationale, design, methods and baseline characteristics of the Asklepios Study. Eur. J. Cardiovasc. Prev. Rehabil. 14, 179191.
  • Unryn BM, Cook LS, Riabowol KT (2005) Paternal age is positively linked to telomere length of children. Aging Cell 4, 97101.
  • Valdes AM, Andrew T, Gardner JP, Kimura M, Oelsner E, Cherkas LF, Aviv A, Spector TD (2005) Obesity, cigarette smoking, and telomere length in women. Lancet 366, 662664.
  • Weiss N (2005) Mechanisms of increased vascular oxidant stress in hyperhomocysteinemia and its impact on endothelial function. Curr. Drug Metab. 6, 2736.
  • Xu D, Neville R, Finkel T (2000) Homocysteine accelerates endothelial cell senescence. FEBS Lett. 470, 2024.
  • Von Zglinicki T, Saretzki G, Docke W, Lotze C (1995) Mild hyperoxia shortens telomeres and inhibits proliferation of fibroblasts: a model for senescence? Exp. Cell Res. 220, 186193.